Bayesian Models: Revolutionizing Dairy Genetics and Breeding Strategies

Bayesian Models: Revolutionizing Dairy Genetics and Breeding StrategiesDairy cattle breeding programs are increasingly confronting a category of traits that defy simple statistical treatment: health phenotypes with skewed or binary distributions, reproductive outcomes that are weakly heritable, novel sensor-derived measurements from mid-infrared spectroscopy or automated milking systems, and sustainability-relevant traits such as methane emissions and feed efficiency that are expensive to record at scale. As genomic reference populations grow and whole-genome sequence data become more accessible, there is a pressing need for statistical frameworks that can handle distributional complexity, incorporate prior biological knowledge, and scale to the high-dimensional marker sets that modern evaluations demand.Bayesian methods — from threshold and hurdle models for categorical health and fertility traits, to the Bayesian alphabet approaches (BayesA, BayesB, BayesCπ, BayesR, BayesRCO) for genomic prediction, to network-based frameworks for dissecting genetic architecture — are uniquely positioned to meet these demands. Recent advances in integrating functional annotation priors from large-scale genomics initiatives, variational inference algorithms for large-scale computation, and multi-trait formulations that jointly model correlated production and functional traits are expanding what Bayesian evaluation can achieve in practice. Yet the gap between methodological development and routine application in national and international breeding programs remains wide.This Research Topic invites contributions that advance Bayesian statistical methodology for dairy cattle genetic evaluation and apply these approaches to phenotypes of current breeding relevance. Themes of interest include:o Genomic prediction under Bayesian variable selection — BayesR and related approaches, performance across trait types, reference population sizes, and marker densitieso Threshold and hurdle models for health and reproductive traits — subclinical mastitis, lameness, ketosis, fertility disorders, and survival traitso Multi-trait Bayesian models — joint estimation of genetic correlations, selection index construction, and handling of missing phenotypes across environmentso Multi-omics integration and high-density data — Frameworks incorporating transcriptomics, metabolomics, and microbiomics into genomic evaluation; integration of imputed sequence-level variants, rare variant weighting, and functional biological priors (e.g., eQTLs) o Across-breed, multi-population, and genotype-by-environment (G x E) inference — Transferability of genomic predictions, informative priors from related breeds, Bayesian reaction norms for climate resilience (e.g., heat stress), and modeling macro and micro-environmental sensitivity o Bayesian networks and systems biology — Causal inference models, structural equation modeling, and multi-layered omics networks to dissect the genetic architecture of complex dairy traits. o Novel phenotypes from sensor and spectroscopy data — Bayesian frameworks for MIR-predicted milk composition, methane proxies, and automated monitoring traitso Computational scalability — variational Bayes, Markov chain Monte Carlo efficiency, and approximate methods for large-scale national evaluationso Genetic diversity and inbreeding — Bayesian estimation of effective population size, runs of homozygosity, and genomic inbreeding in selection contextso Biomarker identification based on Bayesian approaches Topic editors: Juliana Petrini, Gerson Barreto Mourão Source: https://www.frontiersin.org/research-topics/82355/bayesian-models-revolutionizing-dairy-genetics-and-breeding-strategies